DIAGNOSIS AND MANAGEMENT OF BRAIN METASTASES

Susanne M. Arnold MD

Roy A. Patchell MD

¹ Division of Hematology and Oncology, Departments of Medicine, Surgery (Neurosurgery), and Neurology, University of Kentucky Chandler Medical Center  and the Multidisciplinary Lung Cancer Program, Markey Cancer Center, Lexington, Kentucky

Brain metastases are a common complication of systemic cancers, occurring in more than 100,000 patients per year. Metastases to the brain are the commonest intracranial tumors, outnumbering primary brain tumors by at least 10 to 1. In the 1980s and 1990s, significant advancements were made in the diagnosis and treatment of brain metastases. Although the development of brain metastases still usually indicates a poor prognosis for the patient, it is now possible to reverse most of the symptoms of brain metastases and improve significantly a patient's quality and length of life.

FREQUENCY

The frequency of brain metastases has increased over time, likely as a result of modern neuroimaging techniques and more precise autopsy studies.Older estimates suggested that metastases constituted 10% to 15% of all intracranial tumors^[38] ; however, current figures estimate their frequency at 20% to 40%. The frequency of brain metastases may increase as improved survival in cancers that frequently metastasize to the brain occurs.

The origin of the primary tumor is associated strongly with the frequency and pattern of intracranial spread . In adults, the commonest sources of brain metastases are the lung, breast, gastrointestinal tract, genitourinary tract, and skin (malignant melanoma). Cerebral metastatic disease in children is less frequent than in adults, with four retrospective studies estimating the occurrence to be 6% to 10%. The commonest childhood solid tumors that metastasize to the brain are neuroblastomas and a variety of sarcomas, including embryonal rhabdomyosarcoma, Wilms' tumor, Ewing's sarcoma, and osteogenic sarcoma. In children older than 15 years, germ cell tumors have the highest incidence.^] The clinical presentation and neurologic manifestations of cerebral metastases in children are similar to those seen in adults, and the approach to diagnosis and treatment is the same.

Of patients, 5% to 10% may present with acute neurologic symptoms caused by hemorrhage into the tumor or cerebral infarction from embolic or compressive occlusion of a blood vessel.^[19] Hemorrhage into a metastasis is particularly common with choriocarcinoma and melanoma. The signs and symptoms of cerebral lesions are often quite subtle, however; brain metastases should be suspected in all patients with known systemic cancer in whom new neurologic findings develop.

DIAGNOSIS

The best diagnostic tests for brain metastases are contrast-enhanced MR imaging and (to a lesser extent) CT. If the clinical history is typical and lesions are multiple, usually there is little doubt surrounding the diagnosis. It is important, however, that metastases be distinguished carefully from primary brain tumors, abscesses, cerebral infarction, and hemorrhages. Imaging findings that favor metastases include gray-white junction location, relatively smooth margins, a small tumor nidus with a large amount of associated vasogenic edema, and the presence of multiple lesions.It is crucial to identify accurately patients with single metastases whose subsequent management and prognosis may be different.

Contrast-enhanced MR imaging is more sensitive than enhanced CT (including double-dose delayed contrast) or unenhanced MR imaging in detecting lesions in patients suspected of having intracranial metastases. Although T2-weighted MR imaging sequences are sensitive in showing vasogenic edema as areas of increased signal intensity, not all metastatic lesions have sufficient edema to be identified. Even when using contrast MR imaging for the diagnosis of single brain metastases, one study revealed a false-positive rate of 11%.Other diagnostic tests, such as arteriography or biopsy, may be needed to establish the diagnosis firmly. Stereotactic biopsy has become the most convenient and safe approach to obtaining a tissue diagnosis. In addition, newer modalities allow for biopsy and treatment (resection or irradiation) during the same session.

When a brain mass is discovered on MR imaging and there is no prior history of cancer, it is difficult to know how far to pursue investigation. In most cases of brain metastases, the primary tumor resides in the lungs or has metastasized there before disseminating to the brain. More than 60% of patients with brain metastases have a mass apparent on chest radiograph. When the chest radiograph fails to show a mass, CT of the chest may show a lesion and suggest the cause of the neurologic disorder. A CT scan of the abdomen occasionally shows an unsuspected renal or colon cancer. Further search for a primary tumor is almost never fruitful, without positive features in the patient's history or localizing signs on physical examination to suggest a specific primary tumor. When there are no identifiable lesions found in the lung, the mechanism of metastatic spread to the brain may be by a patent foramen ovale (paradoxical embolus), tumor filtration through the lungs without lung metastasis, or dissemination through Batson's vertebral venous plexus.

TREATMENT

Newly Diagnosed Brain Metastases

The treatment of brain metastases includes corticosteroids, radiotherapy, surgical therapy, and radiosurgery. Chemotherapy is useful in some patients with chemosensitive tumors. It is important, however, to distinguish between single and multiple brain metastases when designing a therapeutic plan. Likewise, the extent of systemic disease, neurologic condition, and overall performance status are crucial in treatment planning.

Untreated brain metastases are associated with a median survival time of only about 4 weeks. Nearly all untreated patients die as a direct result of the brain tumor, with increasing intracranial pressure leading to obtundation and terminal cerebral herniation. The survival figure quoted here must be interpreted with caution because the information comes from retrospective studies done before the introduction of neuroimaging with CT and MR imaging. Also, patients who receive no treatment for brain metastases usually have poorer performance status, extensive systemic disease, and poor prognoses, regardless of the presence of brain metastases. It is likely that the average patient diagnosed today would live longer than 1 month, even if untreated.

Anticonvulsants

Seizures occur in about 25% of patients with brain metastases and are the presenting complaint in 10% of patients. Two randomized trials have shown that prophylactic anticonvulsants do not reduce the frequency of the first seizure in patients with newly diagnosed brain metastases. Anticonvulsants should not be given routinely at the time of the diagnosis of brain metastases but rather to patients who actually have had seizures.

Corticosteroids

Almost all patients with brain metastases should be started on corticosteroid therapy at the time of diagnosis. (Patients with small, completely asymptomatic lesions may not need steroids; however, steroids may reduce the side effects of cranial radiation through a reduction in brain edema and are rarely harmful in most patients for short periods.) The beneficial effects of steroids are noticeable within 6 to 24 hours after the first dose and reach maximal effect in 3 to 7 days. The median survival time of patients treated with steroids alone is approximately 2 months, although much longer lengths of survival have been reported.^[ Dexamethasone is the preferred form of corticosteroid because it has minimal mineralocorticoid affect and a relatively low tendency to induce psychosis. More than 70% of patients improve symptomatically after starting steroids.Symptoms reflecting generalized neurologic dysfunction or brain edema respond more consistently than do focal symptoms, such as hemiparesis.

The usual starting dose of dexamethasone is 4 mg four times daily given orally or intravenously. Occasionally, patients require higher doses. With stabilization of symptoms and the completion of more definitive treatment, the dose of dexamethasone should be tapered gradually over several weeks, then stopped to minimize long-term toxicity. About 10% of patients do not tolerate the reduction in steroids and redevelop signs of brain edema. In these patients, the lowest effective dose should be continued indefinitely. Concomitant use of histamine blockers or proton-pump inhibitors to prevent gastritis and peptic ulcer formation is commonplace, but no randomized trials have defined efficacy in this population.

Radiotherapy

Radiotherapy is the treatment of choice for most patients with brain metastases and first was shown to have palliative benefit in 1954. No consensus exists on the optimal radiation dose and schedule for the treatment of brain metastases. Because most patients present with multiple metastases, whole-brain radiotherapy (WBRT) is the commonest form of treatment used in this disorder. The best available data on the effect of dose and schedule for the treatment of brain metastases come from several large-scale multi-institutional trials conducted by the Radiation Therapy Oncology Group (RTOG). These studies have shown no significant difference in the frequency and duration of response for total radiation doses ranging from 20 Gy over 1 week to 50 Gy over 4 weeks. Regimens of 10 Gy in a single dose or 12 Gy in two doses were less effective and no longer are in use.Typical radiation treatment schedules for brain metastases consist of short courses (7 to 15 days) of WBRT with relatively high doses per fraction (150 to 400 cGy/d) and total doses in the range of 30 to 50 Gy. These schedules minimize the duration of treatment, while still delivering adequate amounts of radiation to the tumor.

Accelerated hyperfractionated schemes have been pursued to improve overall survival with varying results. Epstein reported a phase I-II dose-escalation study of hyperfractionated radiotherapy using total doses of 48, 54.4, 64, and 70.4 Gy. No increased toxicity was identified, and the three highest dose arms had a statistically significant improvement in median survival over the lowest dose arm. This study suggested a dose-dependent effect with the use of hyperfractionated radiation in unresected, solitary brain metastasis. A large randomized RTOG trial, studying single and multiple brain metastases, failed to show improvement in overall survival using a hyperfractionated schedule of 54.5 Gy in 34 fractions, compared with a standard regimen of 30 Gy in 10 fractions Additionally, increased focal irradiation to the tumor, or boost dosing, did not increase survival or time to neurologic recurrence when compared with WBRT alone.

Radiation cell sensitizing agents have been used in an attempt to increase tumor cell death. The rationale for using these agents was based on the observation that hypoxic cells (often found centrally in a tumor) are more resistant to the effects of ionizing radiation. Agents such as misonidazole have the potential to increase cell sensitivity to irradiation. None of the radiation cell sensitizers tested to date provide any additional benefit over conventional radiotherapy, however.

WBRT increases the median survival to 3 to 6 months. Data from large retrospective studies have shown that more than half of patients treated with WBRT ultimately die of progressive systemic cancer and not as a direct result of brain metastases.

Retrospective analysis of many patients treated in RTOG brain metastasis protocols has identified patient subgroups that were more likely to respond to WBRT. More favorable outcome is associated with (1) Karnofsky performance scores in the range of 70% or higher, (2) absent or controlled primary tumor, (3) patient age less than 60 years, and (4) metastatic spread limited to the brain. Building on these results, a recursive partitioning analysis was applied to a pooled patient population from several past RTOG studies. The most favorable prognostic group consisted of patients whose Karnofsky scores were 70% or higher, age was 65 years or less, primary tumors were controlled, and no extracranial metastases were present. Performance status at the time of treatment for brain metastases is the most important prognostic factor.

Short-term side effects from radiotherapy are well described. The most serious of these is transient worsening of neurologic symptoms in the initial phase of therapy. During the initial days of treatment, mild symptoms such as nausea, vomiting, headache, and fever are common. This acute reaction may relate to distorted cerebrovascular autoregulation or increased capillary permeability.  Temporary hair loss is predictable, and radiation parotitis and loss of taste may occur with cranial irradiation.

The long-term side effects of radiotherapy are usually not a significant issue in the treatment of patients with brain metastases because of a relatively short survival time of these patients. Reports have suggested, however, that greater than 10% of long-term survivors (>12 months) develop symptoms such as dementia, ataxia, and urinary incontinence. In these patients, imaging studies show cortical atrophy and hyperdense white matter changes. Although the pathogenesis of such alterations is poorly understood, higher dose per fraction and large fractionation schedules play a role. In patients with anticipated long length of survival, a more prolonged course of radiotherapy with smaller doses per fraction should be used. A reasonable schedule for patients with a good prognosis is a total dose of 45 to 50 Gy given in daily fractions of no more than 2 Gy.

Surgery

Surgical therapy is usually not an option for most patients with brain metastases because of the presence of multiple lesions or extensive systemic cancer. In the subgroup of patients whose only metastatic disease is in the brain, however, death is more likely to be due to the brain metastases than to progressive systemic disease. In this group of patients, more aggressive therapeutic approaches, including surgery, are sought for the brain metastases. Several advances have decreased the risks associated with neurosurgery. Safer anesthesia, the widespread use of corticosteroids, the development of modern noninvasive cranial imaging technology, and the introduction of stereotactic approaches all have contributed to a decrease in morbidity and mortality.

There are theoretical reasons that the combination of surgical treatment followed by postoperative radiotherapy may be more effective than WBRT alone. Radiotherapy is most successful when used against small tumor volumes. In larger tumors, radiation is usually effective at the periphery of the tumor where cells are relatively few in number and well oxygenated. In the center, the more hypoxic areas of the tumor, however, radiation may fail to destroy the tumor completely. Although there are documented reports of sterilization of brain metastases by radiotherapy alone,^[21] in most instances, residual tumor remains. Conversely, surgical treatment is most successful in removing large volumes of tumor; however, small numbers of malignant cells may be left behind, especially at the periphery. Rational treatment plans combining surgical debulking and radiotherapy have been developed to overcome the deficiencies of both types of treatment. This combined approach has shown promise in patients with a variety of tumor types

There have been three prospective randomized trials assessing the value of surgical removal of single brain metastases (Table 4) . In a prospective randomized trialperformed at the University of Kentucky, 48 patients with known systemic cancer were treated with either biopsy of the suspected brain metastasis plus WBRT or complete surgical resection of the metastasis plus WBRT. The radiation doses were the same in both groups and consisted of a total dose of 36 Gy given as 12 daily fractions of 3 Gy each. There was a statistically significant increase in survival in the surgical group (40 weeks versus 15 weeks) and an equivalent 1-month mortality of 4% in both groups. The time to recurrence of brain metastases, freedom from death owing to neurologic causes, and duration of functional independence were significantly longer in the surgical resection group. An important finding was that despite screening with contrast-enhanced MR imaging and CT before entry into the study, 6 of 54 (11%) patients with known diagnoses of systemic malignancies were found not to have metastatic brain tumors when tissue was obtained at biopsy or attempted resection. The nonmetastatic lesions consisted of three astrocytomas, two abscesses, and one sterile inflammatory lesion.

Few randomized trials have been reported using radiosurgery. For single metastases, the combined results of many retrospective reports suggest that radiosurgery prevents (or controls) local recurrence of 80% to 90% of treated metastases with about a 5% to 10% risk for radiation necrosis or new neurologic deficits. Auchter et al^[4] reported a retrospective review of the radiosurgery databases of four institutions. The authors found 122 patients who met the same selection criteria used by Patchell in their randomized study (discussed earlier; eg, single brain metastasis; no prior cranial surgery or WBRT; age >18 years; surgically resectable lesion; Karnofsky performance scores >70; and nonradiosensitive histology, such as lymphoma or small cell carcinoma). The patients in the Auchter study received WBRT (median, 37.5 Gy) followed by a radiosurgery boost (median, 17 Gy). The overall local control rate was 86% with an actuarial median survival of 56 weeks and duration of functional independence of 44 weeks. These end points were similar to the surgical arms of the randomized trials of Patchell and Vecht  Despite these apparently promising results, at present, radiosurgery has not been established unequivocally as an effective treatment in the management of single metastases. Prospective randomized clinical trials still are needed and currently are under way to determine the role of radiosurgery in the initial treatment of patients with single metastases and in the management of recurrent brain metastases.

Only one retrospective report has addressed the use of radiosurgery in brainstem metastases. In this study of 26 patients, there were surprisingly few acute symptoms (dizziness, nausea, or seizures). Radiosurgery provided a local control rate of 95%, with a median survival of 11 months, without evidence of herniation or fatal complications in this group of high-risk patients.

Although the addition of WBRT to surgical resection of brain metastases has been established as beneficial, it is unclear whether this benefit can be translated to the radiosurgery setting. In single brain metastases treated with radiosurgery alone, no randomized trials exist comparing WBRT with radiosurgery alone, and the question of cost-effectiveness has not been answered. As stated earlier, radiosurgery is an appropriate substitute for surgical resection of single brain metastases that are technically difficult to resect or in whom patient comorbidities preclude surgery. In terms of multiple brain metastases, the literature is less clear with regard to radiosurgery as a single therapy for brain metastases. A retrospective review of a single institution's experience with radiosurgery with or without WBRT provides some insight into the complexity of the issue. A total of 105 patients with single or multiple brain metastases received radiosurgery with or without WBRT. No change in overall survival was seen, but improved overall failure-free progression (FFP) was noted with the addition of WBRT. In this review and others, there was a significant increase in the risk for recurrent brain metastases and poorer overall intracranial FFP if radiosurgery was used alone. The authors argued, however, that overall survival was not affected secondary to successful salvage treatment with WBRT. Contrasting this argument is work by Suh  which showed a worsening of intracranial FFP when WBRT was delayed rather than given immediately after radiosurgery. No survival difference was seen in this retrospective study. Controversies exist over the role of WBRT in combination with radiosurgery, and further randomized trials are needed to clarify these questions.

The role of radiosurgery in the treatment of multiple metastases has been the subject of three randomized trials. The first randomized trial on the subject was reported by Kondziolka  In that study, 27 patients with multiple brain metastases were randomized to treatment with WBRT alone or WBRT plus a radiosurgery boost. The study was stopped early because the authors claimed to have found a large difference in the recurrence rates favoring radiosurgery. The study used nonstandard end points to measure recurrence. The investigators used any change in measurement of the lesion rather than the standard 25% increase in diameter. No attempt was made to control for steroid use or other factors that might produce small fluctuations in lesion size on MR imaging. Also, a study with only 27 patients lacked the statistical power to support any conclusion, regardless of P values. As a result, this study was uninterpretable.

Preliminary reports of two additional randomized trials examining the use of radiosurgery in the treatment of brain metastases have been published in abstract form.A study reported by Chougule  randomized 109 patients with one to three brain metastases to treatment with radiosurgery alone, radiosurgery plus WBRT, or WBRT alone. There was no statistically significant difference in survival or local control rates among the three treatment arms. Median survival times for the radiosurgery, radiosurgery plus WBRT, and WBRT alone treated groups were 7, 5, and 9 months. This trial suffered from several methodologic problems. The most serious error was that 51 of the patients had surgery for at least one symptomatic brain metastasis before entry into the study. No attempt was made to stratify for previous surgery or otherwise to ensure that surgical patients were distributed equally among the treatment groups. The haphazard distribution of surgically treated patients among the treatment arms weakened the trial and made meaningful statistical analysis impossible.

The RTOG has reported the results of a larger randomized study involving patients with multiple brain metastases. This study (RTOG 9508) contained 144 patients with two or three brain metastases who were randomized to treatment with either WBRT (37.5 Gy) plus radiosurgery or WBRT (37.5 Gy) alone. There was no significant difference in local failure rates in the brain with 21% in the radiosurgery arm and 37% in the WBRT-alone arm (P = .107). There was no significant difference in the length of survival of the two groups (median, 5.3 months for radiosurgery and 6.7 months for WBRT alone). Most noteworthy was the fact that no difference in neurologic causes of death was shown (33% radiosurgery versus 35% WBRT alone), although lower posttreatment Karnofsky scores and steroid dependence were commoner in the WBRT-alone patients. This was a completely negative trial with regard to the major end points of tumor control in the brain, overall survival, and prevention of death resulting from neurologic causes.

The results of RTOG 9508 have forced a major reevaluation of the use of radiosurgery in the treatment of brain metastases. This was the largest and best trial done to date, albeit available only in abstract form, and it failed to show a benefit of radiosurgery in the treatment of multiple brain metastases when radiosurgery was given in the initial management of newly diagnosed tumors. Treatment with WBRT alone is now the treatment of choice in these circumstances. Radiosurgery still may have a place as salvage treatment, however, in patients who have recurrent brain tumors after treatment with WBRT, but this remains to be shown. There have been no definitive randomized trials of radiosurgery in the upfront management of single brain metastases, but RTOG 9508 has a second part evaluating radiosurgery in the management of single metastases that is ongoing, and it is hoped that this will define further the use of radiosurgery in the future.

Interstitial Brachytherapy

The use of interstitial brachytherapy allows delivery of high-dose focal radiation to the tumor, while minimizing the risk for significant radiation exposure to the surrounding normal brain tissue. Permanent and removable implants exist and can be implanted stereotactically or during open surgery. The procedure is limited to relatively small metastases that are located in surgically accessible regions of the brain.

The role of brachytherapy in the primary management of brain metastases has not been determined. Preliminary studies have been inconclusive, and the procedure has potentially serious complications. The major complication is radiation necrosis that may present with the clinical and imaging picture of an expanding mass months after treatment. Sometimes a biopsy specimen is required to differentiate tumor necrosis from recurrence, while steroids and, occasionally, a surgical resection help to reverse the neurologic symptoms secondary to the radiation necrosis. The frequency of this complication varies with the amount of radiation given

At present, brachytherapy cannot be recommended as initial therapy for brain metastases. This technique may offer an additional treatment option for patients with unresectable metastases or patients who have received a prior maximal dose of WBRT.

Chemotherapy

Most systemic chemotherapeutic agents that have proven efficacy against primary cancers have little activity against cerebral metastases from the same cell population. In part this situation is due to inconsistent delivery of drug into the central nervous system. Before the development of brain metastases, an intact blood-brain barrier may prevent dependable dispersion of chemotherapy into the brain parenchyma. Drug entry into the brain depends on molecular weight, lipid solubility, degree of ionization, protein or tissue binding, and local cerebral blood flow. Tsukada  suggested that the blood-brain barrier may increase the incidence of brain metastases after systemic chemotherapy by inducing a relative pharmacologic sanctuary within the central nervous system.

Anatomic barriers play a role, but the blood-brain barrier usually is disrupted at the site of brain metastases and after cranial radiation, and other factors must play a role in the relative chemoresistance found within the brain. Therapeutic levels of many drugs have been shown in the brain, including methotrexate, cisplatin, vinblastine, and carmustine,but do not remain for an adequate period or at high enough concentration to ensure tumor cell death. The drugs appear to concentrate in necrotic areas preferentially, rather than uniformly throughout the tumor, diffusing away rapidly into drug-free normal brain in a sinklike effect.The blood-brain barrier may be intact in one area and not in others within the brain, allowing regional penetration of drugs but not uniform exposure of the whole brain to chemotherapy.

Chemotherapy has been used in the treatment of brain metastases from a variety of primary tumors; however, the results generally have been unimpressive.  In patients with certain highly chemosensitive tumors (eg, breast, small cell lung cancer, germ cell tumors) more effective responses have been seen^[ ; however, chemotherapy is not the primary therapy for these patients and is seldom the only therapy. In a study of 116 patients with melanoma, breast, and lung cancers and brain metastases, cisplatin and etoposide were given every 3 weeks until disease progression.Patients with breast and non-small cell lung cancer had a response rate of 30% and 37%, whereas melanoma patients showed no benefit from therapy. Although it seems that chemotherapy penetrated the central nervous system well in one third of patients, it remains to be seen what the role of chemotherapy will be in the treatment of brain metastases.

Another area in which chemotherapy has been studied is in combination with WBRT for brain metastases from non-small cell lung cancer.The only randomized trial showed an improved response rate (in the brain) when chemotherapy was added to radiation versus radiation alone.In other nonrandomized trials, response rate was improved when compared with historical controls using multiagent, cisplatin-based chemotherapy, with intracranial response rate ranging from 30% to 75%.No neurologic sequelae were identified in the small subset of patients who survived for longer than 18 months in one of these studies.Although radiosensitization appears to be no more toxic than radiation alone, these are phase II studies, and it is unclear whether there is a survival advantage over WBRT alone. The concept of chemosensitizing brain metastases to radiation is intriguing; however, this therapy is experimental and cannot be recommended without supporting data from randomized trials.

At present, a reasonable use for chemotherapy in brain metastases would be in patients with small, asymptomatic tumors from primaries that are known to be chemosensitive. If progression occurs with the patient receiving chemotherapy alone, more definitive treatment with surgery, radiosurgery, or radiotherapy must be given.

RECURRENT BRAIN METASTASES

Brain metastases often recur, and central nervous system progression may be accompanied by systemic tumor progression and a decline in Karnofsky performance score. In general, the same types of treatment used for newly diagnosed brain metastases are available for recurrent tumors. The type of previous therapy may limit the therapeutic options available at recurrence, however, and the development of radioresistance is common.

Commonly, patients with recurrences already have been treated with WBRT, and this limits the amount of subsequent radiation that can be given safely. The amount of additional radiation that can be offered is usually in the range of 15 to 25 Gy, a dose usually inadequate to control tumor growth. Several uncontrolled studies have found no meaningful increase in survival time and marginal improvement in neurologic symptoms in patients who underwent further radiotherapy after the recurrence of brain metastases Because these studies included a relatively heterogeneous group of patients (some with poor Karnofsky performance score, extensive disease, or radioresistant tumors), it is difficult to cull out subgroups who would benefit from reirradiation. It is reasonable to reirradiate patients who showed an initial favorable response to radiotherapy, had a longer disease-free interval, and who remain in good general condition when the cerebral recurrence develops. Even in this favorable subgroup, however, only 42% of patients showed symptomatic improvement, and the median survival after reirradiation was 5 months. Despite such relatively poor results, additional radiotherapy frequently is the only treatment option for patients with recurrent disease.

Conventional surgery for recurrent tumors is an option in patients who have a single recurrence and a good performance status. Sundaresan  reported a series of 21 patients who were treated with craniotomy for their initial brain metastases and who underwent a second craniotomy for recurrence. After the second operation, two thirds of the patients experienced neurologic improvement, and the median survival after operation for the recurrence was 9 months. A second retrospective study examined 48 patients treated with reoperation for recurrent brain metastases and revealed a median survival of 6.7 months after reoperation. In another report of 109 patients with non-small cell lung cancer and recurrent brain metastases, 32 had surgery for their recurrences and survived longer than patients with recurrence who did not undergo surgery.^[2] In all of these studies, the patients were a select group with relatively little systemic disease and a single recurrent metastasis.

It appears that patients who undergo surgical resection for recurrent brain metastases after initial WBRT fare poorly. A study from the Memorial Sloan-Kettering Cancer Center examined this population and reported a median survival after surgical resection of 5 months. When compared with patients who had had surgery plus WBRT as initial therapy for brain metastases, this group had a poorer overall survival. Usually, additional radiotherapy is not given after operation for a recurrence; however, brachytherapy with implantation of removable radioactive sources has been tried in a few patients with incomplete resections. The value of brachytherapy for recurrent metastases is unclear.

Stereotactic radiosurgery has been used to treat recurrent brain metastases. Radiosurgery has the theoretical advantage of being able to deliver large doses of additional radiation to small areas of the brain. Several series of recurrent tumors treated with radiosurgery have been published. The treated lesions were apparently controlled, with a decrease in size or stabilization post treatment and an increase in survival. These patients were a highly select group who had small recurrent tumors and limited systemic disease. Further studies are needed to determine the true value of stereotactic radiosurgery in the management of recurrent brain metastases. Chemotherapy has been used in recurrent tumors, but its benefits, if any, have not been determined.

FOLLOW-UP

There is no set standard for the follow-up of brain metastases after treatment. MR imaging or CT scans are indicated at any time after therapy that patients develop new neurologic symptoms. How frequently asymptomatic patients need follow-up scans is controversial. For patients treated with surgery, a contrast MR imaging scan should be performed within 5 days after surgery to detect residual disease. This scan is especially important if foregoing postoperative radiation therapy is being considered. If residual disease is present, patients should be given WBRT. For all patients treated with WBRT, follow-up imaging should be obtained at regular intervals after treatment. In general, it takes about 6 weeks after WBRT for a definite change to be evident on MR imaging, and patients usually do not need scans immediately after the completion of radiotherapy. A reasonable schedule of follow-up scanning would be to obtain a scan 3 months after completion of last therapy (either WBRT or surgery), then about every 4 months for the first year after treatment. The length of time between scans can be stretched out gradually so that asymptomatic patients are scanned only once per year.